Gainclones and fans, a bad mix ?

[Selexus]

Hi Guys and Gals,

I'm finalising the design of my GC chassis using BrianGT's boards, and have realised that i probably won't get enough passive cooling and will have to add some fans.
I've read several comments with regards to interference from fans but i don't think i've noted anyone that has done a proper test.
Unfortunatly i'm not in a position to do such a test yet so......could some one do one for me ?

Something along the lines of a 40mm pc fan or similar about 50mm ( 2 ") from the chip with a 5mm piece of aluminium seperating them.

I'm going to some considerable trouble to route power, input and output wires away from each other. The last thing i want is to sit down for an evening with Mr Satriani and have Joe Papst buzzing away in the background .

Why is nothing ever easy ?

[tiroth]

I have built low power gainclones with nothing more than a Celeron heatsink for a stereo pair. If the application is not demanding, don't give up hope without testing.

I would never use a fan simply because the mechanical noise alone would be unacceptable.

[TwoSpoons]

The trick with fans is to use a large diameter fan, running slow. This cuts the acoustic noise dramatically. The other good trick is to use a variable speed fan, controlled by a thermistor on the heatsink. So at low volumes, when there's not much heat dissipation the fan runs slow, and at high volumes the fan speeds up - but you wont hear it because of the loud music .

As for electrical interference - use an AC fan if you don't want speed control, or a separate DC supply just for the fan for DC fans. Magnetic coupling can be over come by mounting the fan on the case, sucking air in, then venting out past the heatsink with some simple ducting. This keeps the fan away from the amplifier electronics.

I'm going to be fan cooling my 5 ch 100w/ch GC using 2x 80mm speed controlled DC fans, mounted on the case.

[Oborous]

Hiya,

Just thought I'd offer some alternatives. I've been looking into this myself, as I'm a complete noob to Audio, but do the Computer thing to a high level of anal-retentiveness.

Have you considered just using a large thermal mass? (i.e. big chunk of metal). Heat pipes could also be used to move the load to another area where you could get the surface area to dissapate it. Vents would also be beneficial here.

Or the real suggestion;

Ducting, so very simple, move the noisy component away from the sensitive components. Since fields dissapate at Log-2 attenuation, every bit of distance really helps.

I haven't done any emperical testing on fans (I would suggest SilenX or Papst for high quality, high performance fan solutions), but have plenty of experience with the crappy grade of fan and interference and noise at frequencies that just pierce through any other noise.

I'd suggest the 60mm or 80 mm fan, better performance spec's.

[breez]

Use a bigger fan at lower rpm to achieve same airflow with reduced noise. If you have space that is.

[Vikash]

If you have space that is.

If you have the space then use a bigger block of metal. What size heatsink are you currently using that you think is inadequate?

[Selexus]

Originally posted by Vikash

If you have the space then use a bigger block of metal. What size heatsink are you currently using that you think is inadequate?

The actual GC board will be housed on a piece of aluminium bar that is 60mm in diameter with a large section (pocket) milled out for the board to sit in. Probably equivalent to a 25mm x 50mm rectangular bar in mass . An acrylic tube is then slid over the bar to enclose the board but leave it visible. Brian did a sterling job on designing them so i figured they should be seen not hidden away. There are four of these which house the 4 channels and are connected to 2 40mm copper tubes running front to back.
Kinda like this if you view it from above



The fans are to be housed in the ends of the bars that house the boards and draw air over the boards.
In the big gap in the middle are 2 circular "cans" milled from ali and joined to form a figure 8 that house the troids.

So you see large fans are out.
I dont need a lot of air, just enough to do the job
Hope that all makes sense.

Do fans stall if you run them with to low a voltage ?

[mateo88]

If you use an additional (small) transformer and rectifier you'll be fine. I've done this in several amps and have had no interference at all from the fan. The fan really doesn't have to be running fast at all; you'd be surprised at how little air movement it takes to cool the heatsink well.

[jackinnj]

Originally posted by mateo88
If you use an additional (small) transformer and rectifier you'll be fine. I've done this in several amps and have had no interference at all from the fan. The fan really doesn't have to be running fast at all; you'd be surprised at how little air movement it takes to cool the heatsink well.

Matt makes a good point -- air will move itself at around 32 linear feet per minute (just think how slow you have to walt to cover 32 feet in one minute). A 12 V fan run at 7 volts will move a decent amount of air with few decibels of audible noise.

Getting the heat sink fins so that they are perpendicular to the ground helps (if you aren't going to be moving air.) If the fins are parallel to the ground they just heat each other up.

I found that with my bridged amp the heat sink would go over 200 degrees fahrenheit without the fan, with the fan it was around 105 -- this when it was putting out over 100 watts.

with all due respect to Douglas Self, the distortion was higher when the devices were at their "distress" point.

[karma]

try it without. u can add one later if its to hot. witch likelly will not be the case

just leave the top open so it can vent . and u can see brians work